Improved Monte Carlo Tree Search-based approach to low-thrust multiple gravity-assist trajectory design

Abstract

In the preliminary design of low-thrust trajectories with multiple gravity-assist phases, the selection of the celestial bodies to be considered and the sequence of optimal maneuvers have a significant impact on the overall mission performance. This aspect makes the preliminary transfer trajectory analysis a very complex task. This Short Communication presents a novel method to design heliocentric multiple gravity-assist transfer trajectories for a spacecraft with a low-thrust propulsion system, where the total velocity variation is minimized. In this context, a Bézier shape-based method is used to describe the generic arc of a propelled trajectory, while an extension of the Monte Carlo Tree Search algorithm is used for a rapid analysis of the gravity-assist maneuver sequence. In particular, the proposed approach effectively solves the problem of selecting a suitable balance parameter in the Monte Carlo Tree Search-based routines, which makes the method especially useful in a preliminary mission phase. The effectiveness of the proposed approach is shown by simulating an interplanetary transfer towards the outer regions of the Solar System.